In addition to their vital roles in normal physiology, recombinantly produced cytokines and antibodies constitute a large, critical family of drugs used in the treatments for cancer, HIV, and a host of hematological and immune disorders. However, these molecules are often difficult to produce in mass quantities, suffer from poor folding and aggregation due to the presence of disulfide bonds, and are relatively unstable as therapeutic formulations. Recently, alternative scaffolds with superior biophysical properties to cytokines and antibodies have been used in directed evolution experiments to engineer high-affinity binders to targets of interest. However, cytokine and antibody therapeutics can serve as either agonists or antagonists, and current directed evolution methodologies only select for target binding, not consequent agonism or antagonism. Here, we propose to exploit the modular fold of an alternative scaffold - the designed ankyrin repeat protein - to develop novel directed evolution approaches for isolating new agonists and antagonists. The goals of this R21 proposal are to: 1) engineer anykrin agonists that are mimetics of interleukin-12, a potent anti-tumor cytokine, and 2) create site-specific ankyrin inhibitors of matriptase, a cell-surface protease tha is implicated in a variety of epithelial-derived tumors. More generally, these methods will not only enable the production of new therapeutics, but will also provide unique insights into the basic requirements for specific, high-affinity protein recognition.